Rather than seeing AI as a threat, employers can flip the script by harnessing its capabilities to monitor and enhance employee wellbeing.

Australia is currently one of the most stressed-out countries in the world. Last year, 48 per cent of Australians reported high levels of stress at work, according to a report by Gallup. This makes them the second-most stressed workers globally (alongside New Zealand).

Some reports suggest the proliferation of artificial intelligence (AI) will exacerbate this issue by depersonalising the employee experience. 

However, rather than seeing AI as a threat to employee wellbeing, some are embracing it as an opportunity to provide more comprehensive wellbeing strategies. 

Chatbots as digital sounding boards

It might be easy to assume that most workers would prefer to discuss their wellbeing with a human rather than a machine. However, research shows employees have warmed to AI’s potential benefits for mental health. 

In a 2021 study by Oracle and Workplace Intelligence, 75 percent of employees said AI had helped their mental health at work. 

The top benefits cited were access to useful information, decreasing workloads and helping to prioritise tasks.

Even more intriguing was the finding that only 18 per cent of people would rather speak to humans than robots about their mental health. Respondents said robots provided a judgement-free zone, an unbiased sounding board and quick answers to their sensitive health-related questions.

To help employees access safe digital spaces, some organisations are turning to AI-powered chatbots that provide resources, coping strategies and referrals to mental health professionals. 

AI-powered wellbeing platforms can offer wellbeing check-ins, personalised advice and a neutral space to discuss problems, and act as a 24-hour support to be utilised as a triage system ahead of expert human intervention. 

“If we’re tracking our steps and all sorts of other things, why not track how our body is handling stress at any given time?” – Meg Price, co-founder, Noa Coach

This becomes especially useful in rural and remote areas where there is a lack of qualified mental health professionals. For example, in outback South Australia, there are approximately 19 psychologists for every 100,000 people, according to the Australian Institute of Health and Welfare.

“If I’m struggling, I want in-the-moment help because I want to chat to someone and sort myself out,” says Meg Price, co-founder of Noa Coach. “But I also want something that helps me build healthy habits and wellbeing long-term.”

Although AI can be an invaluable tool for employees who are reluctant to discuss their emotions with a human, Price stresses that it should complement, not replace, human support. 

“It’s not one or the other,” she says. “It should be a helpful tool that allows you to find the words to speak to the people you need to speak to.”

If users score poorly on wellbeing checks or their issues persist, AI-powered apps should direct them to speak to their manager or a mental health professional, she says. 

Spotting the early warning signs of low employee wellbeing

A common challenge for HR is identifying precursors to wellbeing issues before they escalate into visible struggle. 

“While some progress has been made in developing prevention programs, prevention crucially relies on the reliable detection of specific individuals at risk for depression in the near future, which is currently not possible,” says Dr Eiko Fried, Associate Professor of Clinical Psychology at Leiden University in the Netherlands.

To explore how AI could fill this gap, Fried is leading a five-year project called WARN-D, focused on predicting depression using statistics and machine learning. 

His team is following about 2000 students in the Netherlands over two years, using their smartwatches and phones to gather moment-by-moment data. They hope that, after the project, they’ll be able to reliably predict when depression might occur. 

“The smartwatches capture data including activity, stress and sleep quality/duration, all of which relate to depression,” says Fried.

“We believe that if you truly want to understand a complex, dynamic phenotype like depression, you have to study it as it unfolds over time, in all its complexity.”

Harnessing the power of natural language processing

Some employers might feel uncomfortable collecting personal information from employees, and that’s valid. But Price notes that this data collection is becoming common.

“If we’re tracking our steps and all sorts of other things, why not track how our body is handling stress at any given time?”

AI can also be effective in identifying the early warning signs of stress through natural language processing (NLP). This branch of AI enables computers to comprehend, analyse and synthesise human language.

“We believe that if you truly want to understand a complex, dynamic phenotype like depression, you have to study it as it unfolds over time, in all its complexity.” – Dr Eiko Fried, Associate Professor of Clinical Psychology, Leiden University

The NSW Government’s Centre for Work Health and Safety has used NLP to bolster its wellbeing strategy, partnering with software company Pioneera and its AI bot, Coach Indie. 

“The bot connects to office-based communication platforms to detect early signs of chronic stress and burnout in the workplace,” says a spokesperson for The Centre for Work Health and Safety. “Coach Indie analyses written text, word usage, sentence length and use of emojis and emoticons. If early stages of chronic stress are identified, the bot suggests research-backed [actions] such as breathing exercises.”

Pioneera bases its operations on the CSIRO’s eight AI ethics principles to ensure its use is safe, consensual and reliable. 

Tackling AI bias in employee wellbeing interventions

AI wellbeing tech is commonly designed to be personalised, allowing for tailored support. 

One of the concerns with personalised AI interventions is the potential for bias to creep in. For example, AI software might be trained on data that underrepresents or discriminates against a particular group.

“This requires considerable caution moving forward,” says Fried. 

“However, the statistical models themselves are not to blame for this. It’s the input and how we train them.”

To help overcome these concerns, it’s important to gather input from a diverse range of sources, says Price.

“I was in India recently and I thought, ‘How is AI going to help these people in poor areas of India when it’s been developed in a Western country by people who have no idea of the issues they’re facing?’ So the more people we can bring to that conversation about development, the better.”

As AI continues to find its place in the intricate tapestry of the employee experience, HR has a golden opportunity to lead organisations into a new era of wellbeing support where humans and AI converge to nurture, uplift and inspire the workforce. 

This article first appeared in the October/November 2023 edition of HRM Magazine.

Improve your ability to recognise the signs and symptoms of mental health conditions and learn effective strategies to manage health and wellness in the workplace with AHRI’s short course.